Jul 02 2009
∞
Two Centuries On, a Cryptologist Cracks a Presidential Code - WSJ.com
Cool stuff. Kinda sounds like a variant on columnar transposition with a keyed substitution cipher mixed in.
You know, multiple columnar transposition is still one of the most obnoxious and tedious things to decrypt out there. You can work your way through one or two rounds with digraph analysis (as suggested here), but if you go for a few rounds with a very random key things start to get explosively large. Add in some counter-measures against anagramming (such as adding some random character noise and turning real letters into groups of ciphertext letters) and you still have a cipher that’s usable in the modern world, but still doable by hand.
